Obesity explains approximately 20% of all cancers in the U.S. Our current understanding of the mechanisms involved is only partial, and has not led to effective interventions. As the current obesity epidemic continues unabated, especially among the poor and lower middle class, it is feared that recent progress in lowering cancer rates may be reversed, and health disparities worsened. Markedly different patterns of adiposity exist among the five ethnic/racial populations of the Multiethnic Cohort (MEC) Study (African Americans, Japanese Americans, Latinos, Native Hawaiians, whites). Most surprisingly, we observed striking ethnic differences in the association of BMI with cancer. For example, for breast cancer, the association was strongest, and occurred starting at low levels of BMI (>25 kg/m2), for Japanese Americans and Native Hawaiians, two groups with more central adiposity. In contrast, the association in the other ethnic groups occurred only at high BMI levels (>30 kg/m2). We interpret these differences as supporting an important role for body fat distribution, i addition to amount. Our preliminary data also suggest that considering lifestyle variables and biomarkers in addition to BMI and waist measures improves the size predictions for central adiposity, intra-abdominal visceral fat, and liver fat, as assessed by DXA and MRI. Thus far, there has been no comprehensive study of obesity patterns and their determinants in U.S. ethnic/racial minorities and of their relationships with cancer risk. We propose a new program of research that will: 1) describe the amount and distribution of body fat in 5 major U.S. ethnic/racial populations using whole body DXA and abdominal MRI; 2) identify nutritional, behavioral, biochemical, genetic, metabolomic and gut microbiome predictors of these fat distribution patterns; 3) investigate inter-relationships among these predictors to better understand the causes of obesity; 4) build integrated predictive models for these body fat distributions phenotypes; and 5) test prospectively these predicted phenotypes for associations with breast and colorectal cancer risks in the MEC. A better understanding of the causes of obesity in ethnic/racial minorities may open new avenues for prevention and treatment and help in reducing health disparities.